• Specify each definition statement as 1-80 single-byte characters per line. If a definition statement exceeds 80 characters, enter the backslash (\) continuation symbol immediately before the linefeed code, and then continue the specification on the next line. The continuation line must begin in column 1.
• A comment begins with the hash mark (#). Datareplicator regards any information between a hash mark (#) and the end of the line as a comment. A comment cannot continue onto a subsequent line.
• There are no restrictions on the specification order of operands.
• If you specify the same operand more than once, the last specification in the definition file is effective.
• A parameter name without a parameter value will result in an invalid parameter value error.

• A comment must begin with /* and end with */.
• A comment can continue onto subsequent lines; however, each set of delimiter characters (/* and */) must be entirely on one line.
• To specify a single quotation mark (') in a character string literal, specify two consecutive single quotation marks.
• To use a Datareplicator reserved word as an identifier, enclose it in double quotation marks ("). For details about the Datareplicator reserved words, see Appendix B. Datareplicator Reserved Words.
• Specify a semicolon (;) at the end of the final definition statement.

For details about the format of a duplexing definition, see 5.7 Duplexing definition (source) or 5.12 Duplexing definition (target).

If the source HiRDB is a parallel server and different information is specified in HiRDB's environment variables (such as PDDIR, PDCONFPATH, and SHLIB_PATH) for each server machine, you can specify the operands for the environment variables individually for each such server machine. If you omit the definition for any specific server machine, the source Datareplicator at that server machine will use the environment variables defined for the user environment in which the source Datareplicator's commands are executed. This section explains the format and rules for creating definitions for individual server machines.

For each such definition, the source Datareplicator creates a temporary work file named definition-filename_tmp under HDEPATH during analysis of definition files.

(a) Format for definitions for individual server machines

The following figure shows the format for definitions for individual server machines.

Figure 5-4 Format for definitions for individual server machines

(b) Rules for definitions for individual server machines

The following are the rules for definitions for individual server machines:

• Assemble the definitions starting with the common definition section that begins with commondef, followed by the individual definition sections that begin with nodedef(host-name). For details about the operands that can be specified in the common and individual definition sections, see 5.2 Extraction system definition.
• If the source HiRDB is a parallel server and different information is specified in HiRDB's environment variables for individual server machines, specify the operands for the environment variables for each such server machine in the individual definition sections. Otherwise, you can omit the individual definitions.
• Specify the common definition section at the beginning of the definition file.
• You can omit the string commondef. Datareplicator assumes the beginning of the definition file up to the beginning of the first nodedef as the common definition section, regardless of whether the string commondef is specified. If it detects no nodedef(host-name) string, Datareplicator assumes that the entire information is the common definition section. Note that even if you omit the commondef string, you still must specify the required operands in the common definition section.
• If the same nodedef(host-name) string is specified more than once, the first one specified is effective.
• If the host specified in nodedef(host-name) does not belong to the source HiRDB (the primary system if the system switchover configuration is used), an error occurs during definition analysis.
• Specify commondef and nodedef(host-name) using single-byte lowercase letters.
• Specify in nodedef(host-name) the host name of the server machine being defined.
• Specify for the host name in nodedef (host-name) the target host name for the individual definition from among the host names specified in the -x option of the pdstart statement or the pdunit statement in the system common definition for the HiRDB subject to extraction processing. Do not include any spaces in nodedef(host-name).

If the source HiRDB is a parallel server, you can use the extraction environment definition and the transmission environment definition to define an environment for each back-end server. This section explains the format and rules for creating definitions for individual back-end servers.

For each such definition, the source Datareplicator creates a temporary work file named definition-filename_tmp under HDEPATH during analysis of definition files.

(a) Format for definitions for individual back-end servers

The following figure shows the format for definitions for individual back-end servers.

Figure 5-5 Format for definitions for individual back-end servers

(b) Rules for definitions for individual back-end servers

The following are the rules for definitions for individual back-end servers:

• Assemble the definitions starting with the common definition section that begins with commondef, followed by the individual definition sections that begin with besdef(server-name). Specify in the common definition section the operands whose values are to be common to all back-end servers, and specify in an individual definition section the operands whose values are to be specific to the particular back-end server.
• If all back-end servers use the same definitions, specify only the common definition section.
• Specify the common definition section at the beginning of the definition file.
• You can omit the string commondef. Datareplicator assumes the beginning of the definition file up to the beginning of the first besdef(server-name) as the common definition section, regardless of whether the string commondef is specified. If it detects no besdef(server-name) string, Datareplicator assumes that the entire information is the common definition section.
• You can define all operands in both the common definition section and individual definition sections. If both definition sections are specified for a particular back-end server, the information in the individual definition section takes precedence. If a back-end server has no individual definition section, the information in the common definition section is effective for that back-end server. If an operand is not defined in the individual definition section for a back-end server but is defined in the common definition section, the specification in the common definition section is effective for that back-end server.
• If the same besdef(server-name) string is specified more than once, the first one specified is effective.
• If the server specified in besdef(server-name) does not belong to the source HiRDB, an error occurs during definition analysis.
• Specify commondef and besdef(server-name) using single-byte lowercase letters.
• Specify in besdef(server-name) the server name of the back-end server being defined. Do not include any spaces in besdef(server-name).